From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga07.intel.com (mga07.intel.com [134.134.136.100]) by dpdk.org (Postfix) with ESMTP id B33CE1B148 for ; Wed, 10 Apr 2019 16:02:36 +0200 (CEST) X-Amp-Result: UNSCANNABLE X-Amp-File-Uploaded: False Received: from orsmga001.jf.intel.com ([10.7.209.18]) by orsmga105.jf.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 10 Apr 2019 07:02:35 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.60,332,1549958400"; d="scan'208";a="222234492" Received: from bricha3-mobl.ger.corp.intel.com ([10.237.221.35]) by orsmga001.jf.intel.com with SMTP; 10 Apr 2019 07:02:33 -0700 Received: by (sSMTP sendmail emulation); Wed, 10 Apr 2019 15:02:32 +0100 Date: Wed, 10 Apr 2019 15:02:32 +0100 From: Bruce Richardson To: Aaron Conole Cc: konstantin.ananyev@intel.com, dev@dpdk.org Message-ID: <20190410140232.GD718@bricha3-MOBL.ger.corp.intel.com> References: <20190410134517.63896-1-bruce.richardson@intel.com> <20190410134517.63896-2-bruce.richardson@intel.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.11.4 (2019-03-13) Subject: Re: [dpdk-dev] [PATCH 1/2] acl: remove use of weak functions X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 10 Apr 2019 14:02:37 -0000 On Wed, Apr 10, 2019 at 09:54:02AM -0400, Aaron Conole wrote: > Bruce Richardson writes: > > > Weak functions don't work well with static libraries and require the > > use of "whole-archive" flag to ensure that the correct function is used > > when linking. Since the weak functions are only used as placeholders > > within this library alone, we can replace them with non-weak functions > > using preprocessor ifdefs. > > > > Signed-off-by: Bruce Richardson --- > > lib/librte_acl/meson.build | 7 ++++++- lib/librte_acl/rte_acl.c | 18 > > ++++++++++++++---- mk/rte.app.mk | 3 --- 3 files changed, > > 20 insertions(+), 8 deletions(-) > > > > diff --git a/lib/librte_acl/meson.build b/lib/librte_acl/meson.build > > index 2207dbafe..98ece7d85 100644 --- a/lib/librte_acl/meson.build +++ > > b/lib/librte_acl/meson.build @@ -6,7 +6,7 @@ sources = > > files('acl_bld.c', 'acl_gen.c', 'acl_run_scalar.c', 'rte_acl.c', > > 'tb_mem.c') headers = files('rte_acl.h', 'rte_acl_osdep.h') > > > > -if arch_subdir == 'x86' +if dpdk_conf.has('RTE_ARCH_X86') sources += > > files('acl_run_sse.c') > > > > # compile AVX2 version if either: @@ -28,4 +28,9 @@ if arch_subdir > > == 'x86' cflags += '-DCC_AVX2_SUPPORT' endif > > > > +elif dpdk_conf.has('RTE_ARCH_ARM') or dpdk_conf.has('RTE_ARCH_ARM64') > > + cflags += '-flax-vector-conversions' + sources += > > files('acl_run_neon.c') > > This will also need -Wno-uninitialized (otherwise it will generate > warnings about the search_neon_4 and search_neon_8 functions). > > But I don't like papering over these conversions. I'd prefer instead the > patches I posted at: > > http://mails.dpdk.org/archives/dev/2019-April/129540.html and > http://mails.dpdk.org/archives/dev/2019-April/129541.html > > Are you opposed to merging those? > Nope, not in the least. I'm happy enough to rework this patch on top of those - I'd just had forgotten about them in my rush to get a potential solution out here. I did these up quickly to show how easy it is to remove the need for the weak functions and the subsequent linker "--whole-archive" flag. /Bruce PS: I see your patch 2 does not include the Wno-uninitialized flag, is it not needed in your patch, or just an oversight? From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from dpdk.org (dpdk.org [92.243.14.124]) by dpdk.space (Postfix) with ESMTP id 154DBA0096 for ; Wed, 10 Apr 2019 16:02:42 +0200 (CEST) Received: from [92.243.14.124] (localhost [127.0.0.1]) by dpdk.org (Postfix) with ESMTP id 2E9A51B185; Wed, 10 Apr 2019 16:02:38 +0200 (CEST) Received: from mga07.intel.com (mga07.intel.com [134.134.136.100]) by dpdk.org (Postfix) with ESMTP id B33CE1B148 for ; Wed, 10 Apr 2019 16:02:36 +0200 (CEST) X-Amp-Result: UNSCANNABLE X-Amp-File-Uploaded: False Received: from orsmga001.jf.intel.com ([10.7.209.18]) by orsmga105.jf.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 10 Apr 2019 07:02:35 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.60,332,1549958400"; d="scan'208";a="222234492" Received: from bricha3-mobl.ger.corp.intel.com ([10.237.221.35]) by orsmga001.jf.intel.com with SMTP; 10 Apr 2019 07:02:33 -0700 Received: by (sSMTP sendmail emulation); Wed, 10 Apr 2019 15:02:32 +0100 Date: Wed, 10 Apr 2019 15:02:32 +0100 From: Bruce Richardson To: Aaron Conole Cc: konstantin.ananyev@intel.com, dev@dpdk.org Message-ID: <20190410140232.GD718@bricha3-MOBL.ger.corp.intel.com> References: <20190410134517.63896-1-bruce.richardson@intel.com> <20190410134517.63896-2-bruce.richardson@intel.com> MIME-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.11.4 (2019-03-13) Subject: Re: [dpdk-dev] [PATCH 1/2] acl: remove use of weak functions X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" Message-ID: <20190410140232.vK0mHM26dgQNlANi7Q9hvg4IIO5MDeRFwvE8_iu9Gio@z> On Wed, Apr 10, 2019 at 09:54:02AM -0400, Aaron Conole wrote: > Bruce Richardson writes: > > > Weak functions don't work well with static libraries and require the > > use of "whole-archive" flag to ensure that the correct function is used > > when linking. Since the weak functions are only used as placeholders > > within this library alone, we can replace them with non-weak functions > > using preprocessor ifdefs. > > > > Signed-off-by: Bruce Richardson --- > > lib/librte_acl/meson.build | 7 ++++++- lib/librte_acl/rte_acl.c | 18 > > ++++++++++++++---- mk/rte.app.mk | 3 --- 3 files changed, > > 20 insertions(+), 8 deletions(-) > > > > diff --git a/lib/librte_acl/meson.build b/lib/librte_acl/meson.build > > index 2207dbafe..98ece7d85 100644 --- a/lib/librte_acl/meson.build +++ > > b/lib/librte_acl/meson.build @@ -6,7 +6,7 @@ sources = > > files('acl_bld.c', 'acl_gen.c', 'acl_run_scalar.c', 'rte_acl.c', > > 'tb_mem.c') headers = files('rte_acl.h', 'rte_acl_osdep.h') > > > > -if arch_subdir == 'x86' +if dpdk_conf.has('RTE_ARCH_X86') sources += > > files('acl_run_sse.c') > > > > # compile AVX2 version if either: @@ -28,4 +28,9 @@ if arch_subdir > > == 'x86' cflags += '-DCC_AVX2_SUPPORT' endif > > > > +elif dpdk_conf.has('RTE_ARCH_ARM') or dpdk_conf.has('RTE_ARCH_ARM64') > > + cflags += '-flax-vector-conversions' + sources += > > files('acl_run_neon.c') > > This will also need -Wno-uninitialized (otherwise it will generate > warnings about the search_neon_4 and search_neon_8 functions). > > But I don't like papering over these conversions. I'd prefer instead the > patches I posted at: > > http://mails.dpdk.org/archives/dev/2019-April/129540.html and > http://mails.dpdk.org/archives/dev/2019-April/129541.html > > Are you opposed to merging those? > Nope, not in the least. I'm happy enough to rework this patch on top of those - I'd just had forgotten about them in my rush to get a potential solution out here. I did these up quickly to show how easy it is to remove the need for the weak functions and the subsequent linker "--whole-archive" flag. /Bruce PS: I see your patch 2 does not include the Wno-uninitialized flag, is it not needed in your patch, or just an oversight?